HD 3167 b Moves Where Lava Worlds Keep Air

JWST recorded a 38 ppm eclipse for HD 3167 b, more than 5σ too faint for a dark, bare-rock lava world. The dip is strong evidence that this 0.96-day super-Earth, 154 light-years away in Pisces, has an atmosphere cooling its dayside. The planet now sits in a gap where hotter lava worlds show air and slightly cooler ones look stripped.

A 38 ppm Dip Rules Out Bare Rock

HD 3167 b passed behind its K-type star on June 25, 2025, during a 4.3-hour visit with JWST’s Mid-Infrared Instrument in low-resolution mode. When a planet vanishes behind its star, the lost infrared light is the planet’s own heat. A dark rock with no air should have been a much brighter target.

THE ECLIPSE IN NUMBERS

• Measured depth: The team’s adopted white-light eclipse depth of 38 ppm, with an uncertainty of 11 ppm.
• Bare-rock model: A dark, maximally hot rock at this orbit should have produced 104 ± 3 ppm.
• Brightness ratio: The dayside is only 0.57 times as hot as that maximum, written as R = 0.57.
• Against dark rock: That ratio is 5σ inconsistent with a Moon-like or Mercury-like surface.

Three reductions of the same file, SPARTA, Eureka!, and exoTEDRF, all found a faint dayside. The letter in The Astrophysical Journal Letters, led by University of Chicago graduate student Brandon Park Coy, takes 38 ± 11 ppm as the working depth. Equilibrium temperature is 1786 K, about 1513°C, still hot enough to melt silicate rock on the star-facing face. The planet is tidally locked, so that face never sets.

The Bare-Rock Split Sat Too Hot

Lava worlds are small planets whose daysides can melt rock. Coy’s team sets that threshold near a substellar temperature of about 1500 K, with equilibrium temperatures above about 1100 K. Thermal emission had already split the class. The hottest targets look cooler than a bare rock, which points to thick or reflective air. The coldest look like stripped stones.

Where that change happens was a wide band. GJ 367 b, with a substellar irradiation temperature of 1950 K, appears to lack an atmosphere. TOI-431 b, at about 2600 K, does not match a dark rock. HD 3167 b’s irradiation temperature is about 2500 K, so it was built to sit in that hole. After this eclipse, the paper puts a possible transition between about 2000 K and 2500 K.

That pattern cuts against a simple loss rule. Close-in rocky planets take a hard beating from stellar wind and high-energy photons, and the “cosmic shoreline” idea says they should not keep much air. Most terrestrial planets around small M dwarfs still look like bare rock. JWST even mapped one nearby case, LHS 3844 b, to a dark basaltic surface with no air. Lava worlds are the exception, and HD 3167 b is now the coolest of them with the same cooling signature.

Pisces Hosts a Crooked Four-Planet System

The planet is not a loner. NASA’s archive lists four confirmed planets in the system around a bright K0 V star found by K2 in 2016. Coy’s global fit gives HD 3167 b a radius of 1.6 Earth radii and a mass of 4.84 Earth masses. Bulk density comes out at 6.49 g/cm³, below the roughly 8 g/cm³ expected for an Earth-like mix at this size, which can mean a smaller iron core, a thick envelope, or both.

THE HD 3167 PLANETS

Earlier dynamical work found the inner planet’s path near the star’s equator, while the outer planets move on polar orbits, almost at a right angle. Planet c is a low-density mini-Neptune. Planet b is the lava world. The same fit that sharpened those masses also set the 104 ppm bare-rock prediction that the new eclipse missed.

What Is HD 3167 b’s Atmosphere Made Of?

The dayside emission spectrum is not precise enough to name the gases. Large wavelength bins from 5.06 to 11.77 microns scatter from pipeline to pipeline, and the team warns against reading chemistry out of them. Follow-up with JWST’s NIRSpec is the stated next spectrum, not another MIRI temperature point.

Coy told the University of Chicago that the surface is likely silicate-rich, in the same mineral family as Earth’s mantle. Before this study, the working guess for ultra-hot envelopes was vaporized rock. Density argues against a pure rock-vapor shell, because equilibrium vapor at these temperatures is too thin to puff the radius. The interesting leftover is heavier gas, carbon dioxide, carbon monoxide, or water, sitting over a magma ocean that can both feed the air and swallow it back.

Clouds as Bright as Venus Would Also Fit

An atmosphere can cool a dayside in more than one way, and this eclipse cannot pick a single path. Even full heat sharing with the night side still needs a high effective albedo of 0.72, in Venus’s range. A thin Mars-like shell is ruled out at about 4σ. The data match a thick, highly reflective envelope more than a dark stone.

HOW THE DAYSIDE STAYS COOL

• Heat transport: Gas can move energy to the night side, which rock currents cannot do at this scale.
• Reflection: Clouds can bounce starlight before it heats the ground, raising the effective albedo.
• Absorption: Molecules can hide some of the surface’s infrared light, so the telescope records a cooler brightness temperature.

Molten silicates are expected to be dark, which is why the team prefers air over a shiny rock face. That still leaves clouds versus winds. A September 2026 TESS study led by Coy found lava worlds tend to show high geometric albedos next to hot Jupiters at similar temperatures, which fits reflective silicate clouds as a class trait. Kepler already limited HD 3167 b’s optical geometric albedo to less than 0.8, a ceiling too loose to break the tie.

What’s so surprising is that the closer a rocky planet orbits its star, the harder it should be to have an atmosphere, because it’s bombarded by stellar wind and gets more high-energy photons from the star. But it seems that many of these lava worlds do.

Edwin Kite, associate professor of geophysical sciences, University of Chicago

Kite’s line is the population claim hiding under one eclipse: many of these lava worlds do keep air, even though the shoreline says they should not.

Five Lava Worlds and a Moving Line

HD 3167 b joins four other ultra-hot rocks whose daysides run cooler than a bare surface. Coy counts five terrestrial planets in this ultra-hot set with atmosphere evidence, and this one is the coldest of them. The list is short enough to keep straight, and the new point is the temperature, not a sixth detection method.

WHERE THE SPLIT NOW SITS

TOI-431 b and K2-141 b sit on the atmosphere side of that line with HD 3167 b. Density does not sort the sample as cleanly as temperature does. GJ 1252 b is under-dense and still looks bare, while hotter worlds across a range of densities show the cooling signature. If that holds, JWST should stop treating every “cooler” ultra-short-period rock as an airless default.

Early Earth Had a Magma Ocean Stage

The June 2025 eclipse is the first published result from LAVA LAMPS, JWST GO Program 4818, led by Megan Weiner Mansfield at the University of Maryland. The survey is built around 10 ultra-hot lava worlds and a single question: is there a critical temperature where atmospheres appear. This target was the cooler end of that list. Nine others remain, including GJ 9827 b, TOI-1442 b, TOI-1075 b, TOI-1416 b, TOI-500 b, TOI-1807 b, and HD 20329 b.

WHAT STILL HAS TO HAPPEN

• The rest of LAVA LAMPS: More MIRI eclipses on the nine remaining program planets, to see whether the 2000 K to 2500 K split holds.
• A NIRSpec spectrum: Composition for HD 3167 b is still a blank, and mid-infrared photometry will not fill it.
• Optical phase curves: Reflected light from CHEOPS, TESS, or later missions can separate clouds from night-side heat flow.

The gases are the measurement people keep asking for, because a temperature dip is not a chemical formula. That is the right complaint. Until NIRSpec, HD 3167 b is a cooling signature with two viable engines, winds and bright clouds, and a density that leaves room for volatiles.

Despite how inhospitable they are for life, we’re also interested in studying these kinds of planets because we think early Earth might have looked a lot like a lava world. We think that very early in the solar system’s history, when the terrestrial planets formed, they were extremely hot due to the energy from all of the planetesimal collisions. Earth had what’s known as a magma ocean stage with an entirely liquid surface. This result gives us a window into studying what conditions may have been like in Earth’s first couple of million years.

Brandon Park Coy, graduate student in geophysical sciences, University of Chicago

The planet will not host life. It can still show how a molten surface trades gas with a secondary atmosphere, which is the stage Earth passed through before it had oceans. The useful test is not another headline that it “has air.” It is whether the next nine eclipses keep the split at 2000 K to 2500 K, or smear it out.

Frequently Asked Questions

When Was HD 3167 b Discovered?

Andrew Vanderburg and colleagues found HD 3167 b in 2016 with NASA’s K2 mission during Campaign 8, as a 0.96-day transiting super-Earth around the bright star EPIC 220383386. The atmosphere result is a 2025 JWST eclipse of a planet that had already been weighed and sized for a decade.

Why Use a Secondary Eclipse Instead of a Transit?

Transit spectra work well when a thick hydrogen envelope makes a planet look larger at the wavelengths a gas absorbs. Coy noted that Earth-like or Venus-like shells are so thin that the carbon dioxide transit signal is tiny, while a secondary eclipse reads the planet’s own mid-infrared heat and tests whether the dayside is cooler than a bare rock.

Can Anything Live on HD 3167 b?

No. The equilibrium temperature is 1786 K, the dayside is likely molten silicate, and one face stays in permanent daylight. The science case is the magma-ocean climate of young terrestrial planets, not habitability on this orbit.

What Is the LAVA LAMPS Survey?

LAVA LAMPS is JWST General Observer Program 4818, Looking At Vaporized Atmospheres of LAva/Magma Planets Survey, led by Megan Weiner Mansfield. It targets 10 ultra-hot lava worlds with MIRI eclipses to find whether a critical temperature separates planets with atmospheres from bare rocks; HD 3167 b is the first published planet from that set.